* Corresponding author. minh.ho@sydney.edu.au
The effect of different process configurations on the performance
1and cost of potassium taurate solvent absorption
2Minh T. Ho
1*, Enrique Garcia‐Calvo Conde
1, Stefania Moioli
2and Dianne E. Wiley
1 3 1.School of Chemical and Biomedical Engineering, The University of Sydney, Sydney Australia 4 2. Dipartimento di Chimica, Materiali e Ingegneria Chimica “G. Natta”, Politecnico di Milano, Piazza Leonardo da Vinci 32, 5 I‐20133 Milano, Italy 67
Abstract
8 9The main method of capture of CO2 in industry is the use of solvents for CO2 absorption in post‐
10
combustion capture and the benchmark solvent is monoethanolamine (MEA). However, it presents 11
a few disadvantages such as having a high energy requirement while also being corrosive and toxic. 12
Potassium taurate (K‐Tau) is a solvent with the potential to replace MEA because it has similar 13
reaction rates, high cyclic loading, degradation resistant and most importantly, low energy 14
requirement. 15
The objective of this study was to compare and evaluate the effect of different process 16
configurations on the reboiler duty for the precipitating potassium solvent absorption process. 17 Utilising a baseline potassium taurate process, different process configurations were developed in 18 Aspen Plus. These are a cold rich bypass (CRB) of the rich solvent stream to the stripper and a solid‐ 19 liquid separator. The results show that the modified configurations reduce the reboiler duty of the 20
potassium taurate process by approximately 12% through the reduction in sensible heat and 21 vaporization duty. 22 23 Keywords: Potassium taurate, CO2 capture, Stripper configurations 24
2
1. Introduction
25
Traditionally, chemical absorption by amine solvent is used to remove acid gases (Kohl and 26 Nielsen, 1997). Recently, alternative solvents have been investigated for possible application to CO2 27 including amino acid solvents (Aronu et al., 2013, Lerche, 2012, Majchrowicz, 2014, Majchrowicz et 28 al., 2009, Simons et al., 2010, Vaidya et al., 2010, van Holst et al., 2009). This class of solvent may be 29 advantageous in comparison with traditional amines due to the solid precipitation, which, according 30 to the Le Chatelier’s principle, favours the absorption of carbon dioxide and therefore enhance the 31 overall process. Among several types of amino acids that can be used, potassium taurate is one of 32 the promising solvents. 33
Potassium taurate reacts with CO2 (R1) to form a carbamate ion in solution and a zwitterion
34
which precipitates as a result of its limited solubility. 35
CO K CO aq H O aq ↔ 2KHCO aq (R1)
potassium carbonate potassium bicarbonate
The process is described by two main reactions (Sanchez‐Fernandez et al., 2014, Raksajati et al., 36 2016): 37 CO2 absorption reaction: 38 39
CO 2 OOC R NH aq ↔ ¯OOC R NH COO aq ¯OOC R NH aq (R2)
amine Carbamate Zwitterion
40 Protonated amine (zwitterion) precipitation: 41 42 ¯OOC R NH aq ↔ COOH R NH s (R3) 43
3
According to R2, most of the absorbed CO2 takes the form of a carbamate ion. The CO2
44
absorption reaction implies that for every molecule of absorbed CO2, one carbamate ion and one
45 zwitterion are formed (Brouwer et al., 2005). 46 Fundamental research on the use of potassium taurate as a solvent for the capture of CO2 started 47 with initial publications on studies to assess the equilibrium solubility of CO2 in aqueous solutions of 48
potassium taurate. In a series of experimental works, Kumar et al. (2003a) published the 49
crystallisation characteristics of CO2, the vapour‐liquid equilibrium data (Kumar et al., 2003b) and
50 the reaction kinetics of CO2 and potassium taurate (Kumar et al., 2003c). Following research studies 51 covered the investigation of the physical and thermodynamic properties of potassium taurate (Han 52 et al., 2012, Aronu et al., 2013, Wei et al., 2013, Wei et al., 2014). Various publications extending 53 the knowledge on the solubility (Wei et al., 2013, Wei et al., 2014, Sanchez‐Fernandez et al., 2013b, 54
Sanchez‐Fernandez and Goetheer, 2011), absorption kinetics and reaction rates of potassium 55
taurate in the presence of CO2 (Aronu et al., 2013) were published in the following years .
56
The precipitating nature of potassium taurate requires various adaptations to the traditional 57
absorber‐stripper equipment design employed in non‐precipitating systems (Brouwer et al., 2005). 58
Therefore, employing process equipment capable of handling solid material contributes towards 59
achieving the potential that precipitating systems present to reduce energy consumption and cost. 60
In 2003, Versteeg et al. (2003) developed the first patented enhanced CO2 absorption process
61 intended to handle precipitating solids, which was later named DECAB (Versteeg et al., 2003, Feron, 62 2016). 63 1.1. Objectives 64
In this paper, the objective is to provide a preliminary assessment of the potassium taurate 65
system for post‐combustion capture, as well as to compare and evaluate the effect of different 66
process configurations on the performance of the process. 67
4
2. Methodology
68 2.1. Case studies 69 The techno‐economic modelling of this study has been limited to the recovery section of the CO2 70capture process, as represented in Figure 1. Analyses of the costs of the upstream flue gas 71
scrubbing section and downstream CO2 purification have not been included in the analysis.
72
Similarly, transportation and storage of the captured CO2 has not been included. CO2 compression
73 for transport is not included in the technical or general economic analysis in this study. However, to 74 be consistent with previous studies, the final avoided capture costs and net energy penalty includes 75 this component (CO2 is compressed to 110 bar). 76 77 Figure 1 Boundary of the study; encompassing only the absorption, heat exchange and desorption 78 The case study used for the analysis focuses on post‐combustion removal of CO2 from a 500 MW 79
coal‐fired power plant. The wet flue gas entering the absorption process is assumed to be at a 80
temperature of 313.15 K and a pressure of 1.01 bar, with a CO2 composition of 13% mol, 5% mol O2,
81
7% mol H2O and 75% mol N2. The flowrate of the wet flue gas is 19.6 kmols/s. It is assumed that the
82
flue gas is free of any NOx, SOx and ash contents. The CO2 recovery is assumed to be 90%. The CO2
83
emission from the reference power plant is 0.88 t/MWh. 84
Detailed descriptions of the potassium taurate process can be found in (Moioli et al., 2017, 85 Raksajati et al., 2016, Sanchez‐Fernandez, 2013). 86 Utilising the baseline process simulation originally developed by Moioli et al. (Moioli et al., 2017) 87 and shown in Figure 2, this study adapted the process for the following configurations: 88 Cold rich bypass (CRB) of the rich solvent stream to the stripper ( 89
5
a) Figure 3)– the CO2‐rich solvent stream out of the split, with one stream heated in the
90
cross heat exchanger while the other stream is not heated. The CRB split fraction into 91
the stripper is also varied. 92
b) Addition of a solid‐liquid separator (SLS) (Figure 4) – the CO2 rich solvent stream passes
93
through a hydrocyclone to create a solid rich slurry (underflow) and liquid stream 94
(overflow). The underflow‐to‐total‐flow split ratio (U/T) is set at 0.8 (Raksajati et al., 95 2016). 96 c) Combination of both the CRB and SLS configurations (Figure 5). 97 98 99 Figure 2 Process flow diagram of the Baseline potassium taurate process 100 101
6 102 103 Figure 3 Process flow diagram for the cold rich bypass (CRB) configuration 104 105 Figure 4 Process flow diagram for the hydrocyclone (solid‐liquid separation, SLS) potassium taurate process 106
7 107 Figure 5 Process flow diagram of the combined CRB/SLS potassium taurate process 108 2.2. Process modelling 109
This study uses the model developed by Moioli et al. (2018) for the baseline analysis. Details 110 outlining the development of the model can be found in (Moioli et al., 2017, Moioli et al., 2018). In 111 this model all the electrolyte species not present by default in the Aspen software and the proper 112 thermodynamics of the system have been taken into account. The absorber and stripper were both 113 modelled with the RADFRAC column run in rate‐based mode taking into account the mass and heat 114 transfer resistances in ASPEN Plus®(AspenTech, 2014). 115 In the absorber, the unit is a RadFrac column assumed to operate adiabatically. The unit follows a 116 non‐equilibrium mass transfer rate‐based calculation related to the direct and backward reactions 117
for formation of carbamate and to the forward and backward reactions for formation of 118 bicarbonate where the following set of reactions take place: 119 TAUH H O ↔ TAURINE H O (R4) 2 H O ↔ H O OH (R5) HCO H O ↔ CO H O (R6)
8 CO OH → HCO (R7) HCO → CO OH (R8) CO TAU H O → CARBAM H O (R9) CARBAM H O → CO TAU H O (R10) TAURINE H O ↔ TAU H O (R11) 120 As the precipitation reaction cannot be declared when the RADFRAC simulation is run in the rate‐ 121 based mode, the absorber unit cannot simulate the formation of solids. However, the effects which 122 the formation of a solid phase have on the process have been accounted for by (1) modifying the 123
chemical equilibrium data (2) adding a fictitious flash unit which estimates the amount of solid 124
taurine formed. The full details of this Aspen modelling approach have been detailed in Moioli et al. 125
(2018). 126
The stripper has been modelled as a RadFrac column composed of 20 stages, including the 127
condenser and reboiler, and using a non‐equilibrium mass transfer rate‐based calculation. The 128
stripper operates at a pressure of 1.8 bar. The CO2 rich stream enters the desorption unit at the
129
second stage, while the released CO2 and lean solvent leave at the top and at the bottom stages,
130
respectively. The reflux ratio selected for the condenser is 0.8, while the bottoms to feed ratio has 131
been specified to be 0.986, both specifications on a mole basis. The temperature of the stream 132 leaving the reboiler has been specified to be 120 oC, while the temperature at the top stage of the 133 unit has been specified to be 40 oC . 134 The cross heat exchanger has been set with a temperature difference of 10 oC and a minimum 135
temperature approach of 10 oC . The overall heat transfer coefficient remains constant at 850
136
W/m2K.
9 . 138 Table 1 summarises the properties of absorber and stripper for the Baseline potassium taurate 139 process. 140 Table 1 Properties of absorber and stripper for the K‐Tau baseline process 141
Properties Absorber Stripper
Inlet temperature of gas stream [oC] 40 ‐
Inlet temperature of solvent stream [oC] 40 108
Lean loading 0.27 ‐
Calculation type Rate‐Based Rate‐Based
Number of stages 20 20
Condenser ‐ Partial Vapour
Reboiler ‐ Kettle
Reboiler temperature [oC] ‐ 120
Pressure [bar] 1 1.8
Packing type Mellapak Standard 250X Mellapak Standard 250X
Section diameter [m] 20.7 16.6 Packing height [m] 20 17.6 CO2 lean stream flowrate [kton/hr] 16.5 16.5 142 2.3. Economic assumptions 143
Table 2 summarises the key economic assumptions used in this paper. Further details about the 144
cost assumptions for cooling water, electricity and solvent can be found in Raksajati et al. (2016). 145
Costs for LP steam are plant specific as these are based on the efficiency penalty occurring when LP 146
10 steam is extracted from IEAGHG (2014). In this study, LP steam is assumed to be available at 3.5 bar 147 with no pressure drop between the extraction point and the reboiler. Estimation of the steam cost 148 has been extracted from NETL (2015). 149 Table 2 Key economic assumptions 150 Cost year 2016 Real yearly discount rate 7% Plant lifetime 25 years Construction year 40% completion in year 1 and remaining in year 2 Capacity factor 85% Currency USD Water makeup $1.8/kL Cooling water $0.345/GJ Low pressure steam $0.008/kg Solvent $3/kg Electricity $42/MWh
The capture plant is assumed to be an nth‐of‐a‐kind (NOAK) system. This assumption has a
151
considerable impact on the capture cost assumptions ‐ such as contingencies ‐ as NOAK systems 152
achieve cost reductions due to the experience gained, which reduces estimation uncertainty (Al 153
Juaied and Whitmore, 2009). 154
The overall capital cost has been estimated using Turton’s Module Costing Technique (MCT) and 155
the methodology employed in CCS techno‐economic reports by IEAGHG (Turton et al., 2008, 156
IEAGHG, 2014). The equipment costs estimated for the absorption process include the absorber, 157
stripper, reboiler, cross heat exchanger, coolers and condensers, pumps and blowers, dissolution 158
heater exchanger and hydro‐cyclone (for the SLS case). General equipment such as storage facilities, 159
valves, secondary pumps were estimated as 30% of the sum of the cost of the total purchased 160
11
equipment. Table 3 summarises the breakdown in the capital cost calculations, where A, B1, B2, FP
161
and FM are parameters for each item of equipment, with FP depending on the pressure and FM on
162
the material of construction. 163
Table 3 Itemised cost factors
164 Capital cost element Sum of preceding items Purchase Equipment Cost (PEC) A Bare Erected Cost B B f A, B , B , F , F Project Contingencies C 10% of B Process Contingencies D 30% of B Total Plant Cost E B+C+D Inventory Capital F 0.5% of E Pre‐production G 2% of E Owner’s cost H 7% of E
Total Capital Requirement I E+F+G+H
The operating costs include fixed costs such as maintenance, labour, administration and 165
insurance. Variable operating costs comprise cooling water for the coolers, make up water for 166 losses in the solvent cycle, solvent, electricity for pumps and steam cost for the reboiler (Table 2). 167 Cost of capture is calculated for both the amount of CO2 captured and avoided: 168 169 USD t CO captured NPV of project costs NPV of captured CO Eq. 1 170
12 USD t CO avoided NPV of project costs NPV of avoided CO Eq. 2
171
13
3. Results and Discussion
172
3.1. Variations in the Cold Rich Bypass fraction
173
Figure 6 shows the reboiler and condenser duties with variations in the cold rich bypass split 174
fraction into the stripper i.e. the fraction of the stream out of the absorber that bypasses the cross 175
heat exchanger and is sent directly to the stripper. The cases shown are for where the cold bypass is 176
fed into stage 2 and the CO2 rich stream is fed into stage 10. The change in the split was
177 investigated to see its interaction with the overall process and in particular, the reboiler duty. The 178 split was varied from 1% to 15%. Figure 6 shows that the reboiler duty reduces as the amount of 179 cold bypass increases from zero to 0.12, with the lowest reboiler duty and corresponding condenser 180 duty becoming constant beyond 0.12. Using this split ratio, the stages at which the cold bypass and 181 hot CO2 rich stream are fed into the stripper is also varied (Figure 7). The analysis shows that the 182 lowest reboiler duty occurred when the cold bypass is feed into stage 4, and the CO2 rich stream is 183
feed into stage 16. The subsequent results for the CRB and CRB‐SLS cases are based on this 184
scenario. 185
186
14 187 Figure 6 Reboiler and condenser duties as a function of the percentage of the CRB split fraction selected 188 (cold bypass fed into stage 2, hot CO2 rich stream fed into stage 10) 189 190 Figure 7 Change in reboiler duty at different stages of the stripper at CRB of12% 191 0 10 20 30 40 50 60 70 80 90 370 380 390 400 410 420 430 440 450 460 C ondeser Dut y [M W ] Reb oiler Du ty [MW ]
% Cold Rich Bypass
Reboiler Duty Condenser Duty
396 398 400 402 404 0 1 2 3 4 5 Reboiler duty [MW] Cold bypass into stripper stage CO2rich stream into ‐ stage 10 ‐ stage 16
15
192
3.2. Comparison of performance of different configurations
193 Table 4 Key technical outputs for three potassium taurate cases 194 Process parameter Baseline CRB SLS CRB‐SLS Reboiler duty [MWth] 454 401 398 393 Total heat duty [GJ/t CO2] 6.06 5.54 5.30 5.25 Reboiler duty [GJ/t CO2] 4.50 3.97 3.94 3.89 Dissolution heat exchanger [GJ/t CO2] 1.56 1.57 1.36 1.36 Condenser cooling duty [GJ/t CO2] 0.78 0.25 0.86 0.24 Total cooler duty [GJ/t CO2] 5.73 5.74 5.45 6.01 Cross heat exchanger duty [GJ/t CO2] 5.20 5.22 5.65 5.08 Total pumping duty [MW/t CO2] 1.03 1.04 1.16 1.16 Total energy penalty (excluding compression) [MWe] 115.3 105.5 101.2 100.3 Total energy penalty (including compression) [MWe] 149.3 139.5 135.2 134.3 Total energy penalty (%) 29.9% 27.9% 27.0% 26.9% CO2 rich solvent flowrate out of absorber [kton/hr] 16.8 16.8 20.3 20.3 Solvent working capacity [mol CO2/mol solvent] 0.16 0.16 0.12 0.12 CO2 rich loading [mol CO2/mol solvent] 0.43 0.43 0.42 0.42 Temperature of rich solvent at stripper inlet [oC] 108 111 110 110 Temperature at top of the absorber [oC] 56.7 56.7 51.9 52.9 Boil‐up rate (kgmols/s of water) 3.56 3.11 3.10 3.06
16
Table 4 summaries the key technical outputs for the three potassium taurate cases evaluated. 195
The results show that the Baseline case study has the highest total heat duty, with 75% of the total 196
heat duty attributable to the reboiler, while the remaining comes from the dissolution heat 197
exchanger. In the CRB case, the reboiler and thus total heat duty reduce by about 12% from the 198
Baseline decreasing from 4.5 GJ/t CO2 to 3.97 GJ/t CO2. According to Eisenberg and Johnson (1979),
199
the split configuration maximizes the recovery of heat from the heat lean solvent. This is because a 200
cold bypass configuration means that a smaller rich stream flowrate enters the cross heat 201
exchanger (which can also be heated to higher output temperature) resulting in a lower sensible 202 heat duty. Also, in the CRB case, the reboiler duty reduces because of reductions in the vaporisation 203 duty as shown by the boil‐up rate decreasing from 3.56 to 3.11 kmols/s. A significant reduction is 204 also observed for the condenser duty in this case, decreasing from 0.78 GJ/t CO2 to 0.25 GJ/t CO2. 205
This arises because introduction of the cold bypass into the top of the stripper reduces the 206
temperature of the vapour fraction at the top stage and the corresponding work required by the 207
condenser (Figure 8). The cold bypass acts as a cooling source, condensing some of the rising 208 stripping steam and thus aiding the condenser. An additional benefit of using a CRB is the enhanced 209 desorption of CO2 observed for the top part of the column. This is shown in Figure 9, which presents 210 the K‐values for CO2. The K‐values, represent the vapor‐liquid phase equilibrium values at the stage 211 interface. The desorption of CO2 is substantially enhanced in the upper stages of the stripper due to 212
the drop in temperatures. The results suggest that the top section of the stripper is operating at 213
much higher desorption efficiencies than the other lower stages. 214
17 215 Figure 8 Steam composition and temperature profiles in the stripper 216 217 Figure 9 Vapor‐liquid K‐values for CO2 in each stage of the stripper 218 400 800 1200 1600 2000 Conde nser 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Re boil e r K‐values Stage number Baseline CRB SLS CRB/SLS 0 20 40 60 80 100 120 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Condenser 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Reboiler T [ºC] y,steam [mol/mol] Stage number Baseline y,steam CRB y,steam Baseline T CRB T
18
For the SLS case, the total heat duty is less than both the Baseline and CRB cases, with lower 219
duties arising in the dissolution heat exchanger and reboiler. This is caused by the interactions 220
between the pH of the system and the partial pressure of CO2, which is the driving force of the
221
desorption process. When the CO2 rich stream is sent to the hydrocyclone unit, the supernatant,
222
which is basic as it is rich in potassium counter‐ions, is separated from the slurry. The removal 223
reduces the concentration of the potassium ions and, as a result, the pH of the system also 224
decreases. As outlined in Sanchez‐Fernandez et al. (2013a), a lower pH in the stripper feed 225
translates to a higher partial pressure of CO2, facilitating the CO2 desorption resulting in lower heat
226 of absorption and water vaporization duty. Although a reduction in reboiler duty is observed for the 227 SLS compared to the Baseline, there is no real difference when compared to the CRB case. This is 228 because in the SLS case, due to the presence of the hydrocyclone, the flowrate of the CO2 rich slurry 229 stream is much larger. In splitting the CO2 rich stream, the lean loading of the solvent stream into 230 the absorber increases from 0.27 to 0.3 mol CO2/mol solvent. This is because some of the overflow 231 from the solid‐liquid separator is returned directly to the absorber without being regenerated. As a 232
consequence, the working capacity reduces from 0.16 mol CO2/mol solvent to 0.12 mol CO2/mol. To
233
achieve the fixed recovery rate of 90% CO2, without increasing the size of the absorber, a much
234
larger solvent flowrate is required. Although the lower pH from the slurry stream reduces the 235
absorption and vaporization duty for this scenario, the larger solvent stream flow rate offsets any 236
benefits. The results also show that for the SLS case, there is no reduction in the condenser duty 237
compared to the Baseline case. 238
In the CRB‐SLS configuration, in combining the effects of the previous two configurations, it is 239
expected that the additive effects should result in a configuration with the lowest reboiler duty. The 240
results show a reduction of 14% relative to the Baseline case, and 1.5% relative to the CRB and SLS 241
cases. The lower reboiler duty arises from the reduction in the sensible heat component of the 242
regeneration energy (because the inlet streams to the stripper are 2ºC hotter than in the Baseline 243
case) together with the lower vaporization duty. Additionally, the heat duties of the condenser, 244
cross‐heat exchanger and dissolution heat exchanger for this configuration are also the lowest of all 245
19
cases. The heat duty reductions achieved for the dissolution heat exchanger and cross heat 246 exchanger observed in the CRB‐SLS configuration occur due to the supernatant recycle and are the 247 same as those for the SLS case. Similarly, the reduction in the condenser duty is driven by the same 248 enhanced condensation and lower boil‐up effects as observed in the CRB case. 249 Figure 10 and Figure 11 show the temperature profile and steam mass flowrate in the stripper for 250 the four configurations respectively. The figures show that temperature profile of the CRB and CRB‐ 251 SLS cases varies significantly along the stripper and are largely determined by the cooling effect of 252
the cold rich bypass, where lower temperatures are obtained for three fourths of the stripper 253 column. The results suggest that the operation of the stripper and its process variables are closely 254 dictated by the use of a CRB which influences the reduction in the vaporisation duty, while the use 255 of the hydrocyclone primarily impacts the sensible heat. 256 257 Figure 10 Temperature profile of the liquid phase in the stripper for the four configurations 258 75 85 95 105 115 Condenser 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Re boiler T [ºC] Stage number Baseline CRB SLS CRB/SLS Cold bypass fed here
20 259 Figure 11 Steam mass flowrate profile for each of the four configurations 260 261 3.3. Economic evaluation 262 The capture costs for the different process configurations are summarized in Table 5 and Figure 263 12. The amount of CO2 avoided and avoided cost presented in Table 5 includes the costs and energy 264 consumption for CO2 compression to 110 bar. 265 266 267 268 269 270 5 10 15 20 25 30 35 Condenser 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Re boiler H2O [kg/s] Stage number Baseline CRB SLS CRB/SLS
21 Table 5 Summary of economic results for potassium taurate absorption process 271 Economic parameter Baseline CRB SLS CRB‐SLS Total capital cost for absorption process only [$/kW (Net power)] $1,308 $1,273 $1,455 $1,407 Total capital cost for absorption + compression [$/kW (Net power)] $1,430 $1,392 $1,573 $1,524 Total operating cost for the absorption process [$/kWh] $32 $29 $29 $28 Total operating cost for the absorption process + compression [$/kWh] $37 $34 $34 $33 CO2 avoided for absorption + compression [million t/yr] 1.78 1.84 1.87 1.88 Energy penalty [MWe] 149.5 139.5 135.2 134.3 $/t CO2 avoided (without compression) 61.5 56.9 58.6 57.3
$/t CO2 avoided (with CO2 compression) 76.4 70.9 72.8 71.3
272 Figure 12 Cost breakdown of the four potassium taurate configurations 273 274 0 10 20 30 40 50 60 70 Baseline CRB SLS CRB‐SLS USD/ t avoided Stripp Set up costs Stripp Operating costs Stripp Cooling & heating Stripper Absorber Miscellaneous Blower and pumps Hydrocyclone
22
The economic results show that the configuration with the lowest cost is the CRB. This 275
configuration has costs 5% lower than the Baseline because of the lower energy penalty and 276
subsequent higher CO2 avoided. Although the SLS and CRB‐SLC configurations have lower energy
277
penalties and higher CO2 avoided than the CRB configuration, their overall avoided costs are higher.
278
This is because for these cases, higher total capital costs offset any cost benefits of the reduced 279
energy penalties and lower operating costs. The higher capital costs arise due to additional 280
equipment such as the hydrocyclone and the larger cross‐heat exchanger needed from the larger 281 solvent flowrate. The economic results also suggest that although the CRB‐SLS configuration results 282 in better performance compared to the Baseline, economically it is not competitive. 283 Table 6 presents the change in capital cost for the three process modifications in greater detail, 284 relative to the Baseline case. The results show that for all three configurations, the capital cost of 285 the cross heat exchanger is much higher than in the Baseline, and is particularly so for the SLS case. 286 The increase occurs because the cross heat exchanger unit for these cases requires a much larger 287 surface area relative to the Baseline case. The increase in area needed is the result of the reduction 288 of the temperature driving force of the streams entering the unit. This reduction is substantial in the 289 SLS case, where the cross heat exchanger is 50% larger, with the increase in the heat transfer area 290 requirements caused by the larger solvent flowrates (Table 4). For the CRB case, this increase is 12% 291 and is needed to compensate for the lower flowrate in the cold side. 292 The largest capital cost reductions occur for the condenser and reboiler for the CRB and CRB/SLS 293
cases, however the reverse is true for the SLS where an increase in capital cost relative to the 294 Baseline case is observed. An important feature of the stripper inlet temperature is how it affects 295 the reboiler and condenser duties. The hotter the inlet, the less sensible heat must be provided by 296 the reboiler. However, hotter inlet temperatures lead to higher stripping steam vaporisation rates, 297
which lead to an increase in the condenser duty. This occurs because the stripping steam cannot 298
benefit as much for condensation from a hotter (compared to the baseline case) down‐flowing 299
solvent. This is exemplified in the case of the SLS, where the reboiler and condenser are markedly 300
affected in opposite directions. While the reboiler duty is decreased by 14%, the condenser duty is 301
23 increased by 9%. This trade‐off of duties is unique to the SLS case, as the enhanced condensation of 302 the cases with a CRB negates any vaporisation due to higher temperatures into the stripper while 303 simultaneously leading to lower reboiler duties. The opposite effect in the duties in the SLS case is 304
one instance where hotter temperatures for the CO2 rich stream into the stripper are desired, as
305
the economic results suggest that it is reductions in the LP steam what should be sought, even if 306
this leads to increased cooling water consumption in the condenser. 307
Minor capital cost reductions for the alternate configurations relative to the Baseline case are 308
also observed for the dissolution heat exchanger, coolers, and pumps; primarily resulting from 309 lower solvent flowrates passing through these units. 310 311 Table 6 Changes in the capital cost for the three configuration relative to the Baseline. 312 CAPEX Process block Unit Percentage change relative to Baseline CRB SLS CRB/SLS Heating and cooling Cross heat exchanger 12% 29% 7% Dissolution heat exchanger ‐ ‐8% ‐8% Cooler ‐ ‐4% 2% Stripper Condenser ‐49% 22% ‐51% Reboiler ‐7% ‐8% ‐8% Pumps CO2 rich stream pump ‐ 4% 4% CO2 lean stream pump ‐ ‐10% ‐11% Total CAPEX ‐ 15% 12% 313 Table 7 shows the change in operating cost for the three process modifications, relative to the 314
24 Baseline case. The analysis shows that the main reductions in operating costs for the CRB, SLS and 315 CRB‐SLS cases generally occur because less cooling water is needed in the coolers and condenser, 316 and less steam is required for solvent regeneration in the reboiler. Overall, the operating costs are 317 about 6% to 8% lower than the Baseline, or $3‐4/kWh lower. 318 319 Table 7 Changes in the operating cost for the three configuration relative to the Baseline 320 OPEX
Process block Unit Consumable
Percentage change
CRB SLS CRB/SLS
Absorber Absorber vessel Water makeup ‐ ‐50% ‐50%
Heating and cooling Dissolution heat exchanger LP steam 2 % ‐13% ‐12% Cooler Cooling water 18% ‐5% 5% Stripper Condenser Cooling water ‐62% 9% ‐70% Reboiler LP steam ‐12% ‐13% ‐14% Pumps CO2 rich stream pump Electricity ‐ 4% 4% CO2 lean stream pump Electricity ‐ ‐9% ‐9% Fixed operating costs ‐ 14% 11% Total OPEX ‐6% ‐6% ‐8% 321
3.4. Comparison with MEA
322
Compared to an absorption process using MEA solvent, the energy penalty of the baseline 323 potassium taurate system at 29% is similar to that of a non‐optimised MEA process of 27% (Li et al., 324 2016). As with the MEA process, the largest component of the energy usage in the potassim taurate 325 system is the reboiler (accounting for 75% of the total energy penalty) and the CO2 compressor (at 326
25
23% of the total), while for the MEA process the breakdown is about 58% and 30% respectively (Li 327
et al., 2016). In the literature, typical costs for MEA solvent CO2 avoided range from US$62.0 to
328
US$95.2/tonne CO2 (Li et al., 2016, NETL, 2015, NETL, 2010, Raynal et al., 2011); with the wide range
329
in values being due to different capital cost methodologies, process boundaries and economic 330
assumptions. Using the same economic assumptions as this study, the estimated cost for the 331
baseline MEA process from the study by Li et al (2016) is $72/t CO2 avoided, which is slightly
332
cheaper than the estimated costs for the potassium taurate system in this paper. This difference 333
arises because of the smaller reboiler duty (4 GJ/t of CO2) and solvent flowrate for the MEA system.
334
With regards to the capital cost estimates, for the potassium absorption processes these range from 335
$1,308/kW for the baseline to $1,455/kW for the SLS case. Similar values for capital cost of 336 $1,357/kW have been obtained for a baseline (non‐optimised) MEA (Li et al., 2016). 337 However, the contribution of different process blocks to the total cost is significantly different for 338 the two absorption processes. As shown in Errore. L'origine riferimento non è stata trovata., the 339 largest differences are in the compressor, stripper, heating and cooling process blocks. 340 341 Figure 13 Breakdown in costs for the baseline KTau capture costs and an MEA process 342 0% 20% 40% 60% 80% 100% This study Li et al 2016 % Total Capital C ost Misc. Compressors Pumps and blowers Heating and cooling Stripper Absorber
26
For the MEA process (based on the breakdown in capital cost as detailed in Li et al. (2016), the 343
largest cost component is for the compressor which comprises 40% of the total equipment bare 344 erected cost (BEC). In this study, the compression makes up less than 15%. In contrast, the stripper 345 contributions and heating and cooling blocks are much larger for the potassium taurate system. In 346 this study, a very large solvent flowrate is required due to the small working capacity of the solvent 347 (0.12 mol CO2/mol solvent), which results in very large and more expensive cross heat exchangers 348
being needed. The stripper for the potassium taurate system is also higher and wider at 17.6m x 349
16.6m compared to 7m x 9.5m for the MEA system in Li et al. 2016. In terms of the cost breakdown, 350
the main difference between the two absorption processes lie in the heating and cooling block. In 351
the MEA process, this system accounts for about 4% of the total capture cost, while in the 352
potassium taurate process it accounts for 10%. The larger cost component arises because the lower 353
capacity of the potassium taurate solvent, which is 0.10 mol CO2/mol solvent smaller than that of
354
MEA, leads to a larger solvent flowrate being required. A very large cross heat exchanger is also 355 required, resulting in high costs both in absolute terms and on a per tonne of CO2 avoided basis. 356 The costs estimates presented in this paper for the potassium taurate system are indicative Nth‐ 357 of‐a‐kind values with a 30% error margin. Such Nth‐of‐a‐kind costs will be significantly lower than 358
first‐of‐kind plant costs because technology learning will significantly reduce the capital and 359
operating cost. Furthermore, it should be noted that all costs contain uncertainties due to changes 360
in factors such as exchange rate, fuel costs and labour rates. As a consequence of these 361
uncertainties, if capital costs for the baseline potassium taurate system is 20 % higher than the 362 estimate, the capture cost for the process would likely increase by six percent or $4/t CO2 avoided. 363 Similarly, if the operating costs varied by 20%, through higher fuel or steam cost, the capture cost 364 for the baseline process would increase by almost 15% or $11/t CO2 avoided. 365 366
4. Conclusion
367This paper presents an evaluation of the performance and cost of three alternate process 368
27
configurations for post‐combustion absorption using a precipitating potassium taurate solvent. The 369
Baseline configuration was modified to include a cold‐rich‐bypass and/or solid‐liquid‐separator. By 370
implementing these process changes, the regeneration duty for the process reduces by 371
approximately 12% to 14%, with the flow on effect of lowering operating costs. While a solid‐liquid‐ 372
separator is beneficial for performance, if no changes to the absorber are undertaken, then the 373 increase in capital costs due to the addition of the hydro‐cyclone offsets any cost reductions from 374 lower regeneration duties. The configuration resulting in the lowest capture cost was the cold‐rich‐ 375 bypass, as lower operating costs were realized without the compromise of higher capital costs. 376 In evaluating the different ratios of cold‐rich‐bypass fraction into the stripper, it was found that a 377 split of 0.12 of the cold CO2 rich solvent being sent to the stripper resulted in the lowest reboiler 378
and condenser duties. Further increases in the split ratios were found to make no further 379
improvements in the performance. 380
The techno‐economic analysis of the different process configurations for potassium taurate 381
absorption presented this paper provides a high level assessment highlighting areas of possible 382
performance improvements and cost reductions. In this study, the working capacity and the 383
corresponding solvent flowrates were not optimized. Further studies optimizing the absorber and 384 stripper designs, coupled with investigating heat integration opportunities for these configurations 385 would be worthwhile. Additional experimental studies to obtain more detailed data of fundamental 386 solvent characteristics for the process simulation would also be valuable, as well as investigating the 387
impact of alternate configurations such as absorber inter‐cooling on the absorption and 388
precipitation process.
28
5. References
390 Al Juaied, M. & Whitmore, A. 2009. Realistic costs of carbon capture. Energy Technology Innovation 391 Policy Research Group, Belfer Center for Science and International Affairs, Harvard Kennedy 392 School of Government, Harvard University, Cambridge, MA (United States) 393 Aronu, U. E., Ciftja, A. F., Kim, I. & Hartono, A. 2013. Understanding Precipitation in Amino Acid Salt 394 systems at Process Conditions. Energy Procedia, 37, 233‐240. 395 Aspentech 2014. ASPEN Plus® Guidelines, Burlington, MA, AspenTech. 396 Brouwer, J., Feron, P. & Ten Asbroek, N. Amino‐acid salts for CO2 capture from flue gases. Fourth 397 annual conference on carbon capture & sequestration, 2005. 398 Eisenberg, B. & Johnson, R. R. 1979. Amine regeneration process. Patent No. US 4152217 A1. 399Feron, P. 2016. Absorption‐Based Post‐Combustion Capture of Carbon Dioxide, Woodhead 400
Publishing. 401
Han, X.‐W., Zhou, C.‐R. & Shi, X.‐H. 2012. Determination of specific heat capacity and standard 402 molar combustion enthalpy of taurine by DSC. Journal of thermal analysis and calorimetry, 403 109, 441‐446. 404 Ieaghg 2014. Techno Economic evaluation of different Post Combustion CO2 Capture Process Flow 405 Sheet Modifications. 2014/08. 406 Kohl, A. L. & Nielsen, R. 1997. Gas Purification, Houston, Texas, USA, Gulf Publishing Company, Book 407 Division. 408 Kumar, P., Hogendoorn, J., Feron, P. & Versteeg, G. 2003a. Equilibrium solubility of CO2 in aqueous 409
potassium taurate solutions: Part 1. Crystallization in carbon dioxide loaded aqueous salt 410 solutions of amino acids. Industrial & engineering chemistry research, 42, 2832‐2840. 411 Kumar, P., Hogendoorn, J., Timmer, S., Feron, P. & Versteeg, G. 2003b. Equilibrium solubility of CO2 412 in aqueous potassium taurate solutions: Part 2. Experimental VLE data and model. Industrial 413 & engineering chemistry research, 42, 2841‐2852. 414
Kumar, P., Hogendoorn, J., Versteeg, G. & Feron, P. 2003c. Kinetics of the reaction of CO2 with 415 aqueous potassium salt of taurine and glycine. AIChE Journal, 49, 203‐213. 416 Lerche, B. M. 2012. CO2 Capture from Flue gas using Amino acid salt solutions. PhD Thesis, Technical 417 University of Denmark. 418
29
Li, K., Leigh, W., Feron, P., Yu, H. & Tade, M. 2016. Systematic study of aqueous monoethanolamine 419
(MEA)‐based CO2 capture process: Techno‐economic assessment of the MEA process and 420
its improvements. Applied Energy, 165, 648‐659. 421
Majchrowicz, M. E. 2014. Amino Acid Salt Solutions for Carbon Dioxide Capture. Master Thesis, 422 University of Twente. 423 Majchrowicz, M. E., Brilman, D. W. F. W. & Groeneved, M. J. 2009. Precipitation regime for selected 424 amino acid salts for CO2 capture from flue gases. Energy Procedia, 979‐984. 425 Moioli, S., Ho, M. T. & Wiley, D. E. 2017. Simulation of CO2 Removal by Potassium Taurate Solution. 426 Chemical Engineering Transactions, 57, 1213‐1218. 427 Moioli, S., Ho, M. T., Wiley, D. E. & Pellegrini, L. A. 2018. Thermodynamic modeling of the system of 428 CO2 and potassium taurate solution for simulation of the carbon dioxide capture process. 429 Chemical Engineering Research and Design, 136, 834‐845. 430 Netl 2010. Cost and Performance Baseline for Fossil Energy Plants Volume 1: Bituminous Coal and 431 Natural Gas to Electricity. DOE/NETL‐2010/1397. 432 Netl 2015. Cost and Performance Baseline for Fossil Energy Plants Volume 1a: Bituminous Coal (PC) 433 and Natural Gas to Electricity Revision 3. DOE/NETL‐2015/1723 434 Raksajati, A., Ho, M. T. & Wiley, D. E. 2016. Understanding the Impact of Process Design on the Cost 435
of CO2 Capture for Precipitating Solvent Absorption. Industrial & Engineering Chemistry 436
Research, 55, 1980‐1994.
437
Raynal, L., Bouillon, P.‐A., Gomez, A. & Broutin, P. 2011. From MEA to demixing solvents and future 438
steps, a roadmap for lowering the cost of post‐combustion carbon capture. Chemical 439
Engineering Journal, 171, 742‐752.
440
Sanchez‐Fernandez, E. 2013. Novel Process Designs to Improve the Efficiency of Postombustion 441
Carbon Dioxide Capture. Technische Universiteit Delft,.
442
Sanchez‐Fernandez, E., De Miguel Mercader, F., Misiak, K., Van Der Ham, L., Linders, M. & Goetheer, 443
E. 2013a. New process concepts for CO 2 capture based on precipitating amino acids. 444 Energy Procedia, 37, 1160‐1171. 445 Sanchez‐Fernandez, E. & Goetheer, E. L. V. 2011. DECAB: Process development of a phase change 446 absorption process. Energy Procedia, 4, 868‐875. 447 Sanchez‐Fernandez, E., Heffernan, K., Van Der Ham, L., Linders, M. J., Brilman, D. W. F., Goetheer, E. 448
L. & Vlugt, T. J. 2014. Analysis of process configurations for CO2 capture by precipitating 449
30
amino acid solvents. Industrial & Engineering Chemistry Research, 53, 2348‐2361. 450
Sanchez‐Fernandez, E., Heffernan, K., Van Der Ham, L. V., Linders, M. J., Eggink, E., Schrama, F. N., 451
Brilman, D. W. F., Goetheer, E. L. & Vlugt, T. J. 2013b. Conceptual design of a novel CO2 452
capture process based on precipitating amino acid solvents. Industrial & Engineering 453
Chemistry Research, 52, 12223‐12235.
454
Simons, K., Brilman, W., Mengers, H., Nijmeijer, K. & Wessling, M. 2010. Kinetics of CO2 absorption
455
in aqueous sarcosine salt solutions: influence of concentration, temperature, and CO2
456
loading. Industrial & Engineering Chemistry Research, 49, 9693‐9702. 457
Turton, R., Bailie, R. C., Whiting, W. B. & Shaeiwitz, J. A. 2008. Analysis, synthesis and design of 458
chemical processes, Pearson Education.
459
Vaidya, P. D., Konduru, P., Vaidyanathan, M. & Kenig, E. Y. 2010. Kinetics of carbon dioxide removal 460
by aqueous akaline amino acid salts. Industrial & Engineering Chemistry Research, 49, 461 11067‐11072. 462 Van Holst, J., Versteeg, G. F., Brilman, D. W. F. & Hogendoorn, J. A. 2009. Kinetic study of CO2 with 463 various amino acid salts in aqueous solution. Chemical Engineering Science, 64, 59‐68. 464 Versteeg, G. F., Kumar, P. S., Hogendoorn, J. A. & Feron, P. H. M. 2003. Method for absorption of 465 acid gases. 466 Wei, C.‐C., Puxty, G. & Feron, P. 2014. Amino acid salts for CO 2 capture at flue gas temperatures. 467 Chemical Engineering Science, 107, 218‐226. 468 Wei, S. C.‐C., Puxty, G. & Feron, P. 2013. Amino acid salts for CO2 capture at flue gas temperatures. 469 Energy Procedia, 37, 485‐493. 470 471